Menu
Skills in this repository
SkillsMP has collected 4,114 skills from hiyenwong/ai_collection. Open a skill to review its source and details.
hiyenwong/ai_collectionShowing 40 of 4,114 collected skills.
Methodology for canonical quantization of classical computational primitives (neurons, activation functions, energy-based models) into quantum ML models. Use when designing quantum neural architectures, constructing quantum Hamiltonians from classical energy…
Central Limit Theorem and Sanov's principle for Quantum Neural Network Mixture of Experts (QNN-MoE) — statistical mechanics framework for analyzing parameter fluctuations, convergence, and neural tangent kernel dynamics in quantum neural network ensembles.
Parameter-efficient continuous-variable photonic quantum neural networks for edge medical AI. Room-temperature quantum ML for medical image classification with 40-45% parameter reduction. Covers CV-QNN architecture simplification, barren plateau mitigation,…
Controlled comparison methodology showing continuous-variable (CV) QNNs outperform discrete-variable (DV) QNNs on spatial pattern recognition tasks like wafer-map defect classification.
Geometric Quantum Machine Learning (GQML) toolbox for graph problems — comprehensive characterization of constituents for n-node graphs encoded in n-qubit states. Provides design patterns for quantum graph models including natural classical integration,…
Hybrid quantum-classical neural network methodology for NLP tasks including sentiment analysis and text classification. Uses TF-IDF vectorization plus parameterized quantum circuits, with demonstrated transfer learning advantages over classical baselines.
Quantitative theory for integrability-to-chaos transition in quantum many-body systems via tunable integrability-breaking gates. Use when analyzing OTOC crossover from integrable to chaotic regimes, computing butterfly velocity and front broadening, studying…
Stabilizer state testing and learning under quantum memory constraints. Proves testing complexity is Θ(n-k) with k-qubit memory, learning is Θ(n²/k), and exponential lower bound for purity testing. Identifies coherent quantum memory as the resource enabling…
Multi-agent LLM architecture (QPipe) that autonomously converts natural language requirements into executable quantum application workflows through specialized agents for requirement parsing, formulation, code generation, review, execution, and verification.
Quantum entanglement analysis methodology using logarithmic negativity — proving that logarithmic negativity typically equals exact entanglement cost. Use when quantifying entanglement, analyzing state preparation costs, or studying quantum resource theory.
Mesoscopic linear spectral statistics for random quantum graph ensembles - proves variance coincides with GOE/GUE in large graph limit. Trigger words: mesoscopic statistics, quantum graphs, random graph ensemble, spectral variance, GOE, GUE, Haar measure
Quantum-metabolic bounds on information capacity of noninvasive brain imaging (arXiv:2511.06401)
Metabolic quantum limit methodology for magnetoencephalography - deriving technology-independent bound on brain information capacity using energy resolution and Planck's constant
Noise-aware synthesis methodology for quantum LDPC encoder circuits using two-sided Hamming descent. Enables hardware-aware circuit synthesis for fault-tolerant QEC that accounts for physical noise characteristics.
Partially-Blind Single-Qubit Classification (PB-SQC) methodology for quantum-secured delegated machine learning on untrusted quantum networks. Combines single-qubit classifiers with blind quantum computation to deliver privacy-preserving quantum ML…
Quantum Error Recovery (QER) methodology for permutation-invariant (PI) quantum codes under correlated noise. Uses channel-aware recovery maps with tunable PI code parameters to achieve fidelity beyond noise-independent QEC. Covers CAD code family…
Post-quantum cryptography (PQC) migration framework for Internet of Medical Things (IoMT) with edge-native federated learning security
Post-quantum secure pharmacovigilance data pipeline methodology using ML-KEM-768, ML-DSA-65, HKDF-SHA-256, and AES-256-GCM. Educational prototype for healthcare data pipeline security in the post-quantum era. Covers component architecture, file format…
Quantum Convolutional Neural Network (QCNN) methodology for surrogate modeling of complex physical systems. Uses quantum convolutional and pooling layers with Hamiltonian-inspired encoding, benchmarked across simulators and real quantum hardware with error…
Hybrid quantum-classical transfer learning methodology showing 15 percentage point accuracy improvement on spam classification (66%→81%) when transferring from COVID-19 sentiment analysis. Demonstrates enhanced generalization of QML models through transfer…
Scalable on-hardware training of Quantum Neural Networks for clinical data imputation methodology - demonstrates practical quantum machine learning for handling missing data in clinical datasets.
Methodology for stress-testing variational quantum neural networks on complex physical datasets, with emphasis on hyperparameter optimization, expressivity enhancement, and classical baseline comparison.
LLM-based multi-agent architecture for autonomous quantum application generation from natural language requirements. Use when building agentic systems for quantum software engineering, automated quantum code generation, NL-to-quantum workflows, or quantum…
Benchmark for measuring API drift in LLM-generated quantum code across successive SDK versions. Evaluates version fidelity, cross-version compatibility, failure modes, and documentation-guided repair. Instantiated with Qiskit v0.43, v1.3, v2.0. Activation:…
Quantum convolutional autoencoder (QCAE) for reconstruction-based anomaly detection using QCNN architectures - semi-supervised training on normal samples with reconstruction error as anomaly score.
Contraction and expansion values as monotone sequences refining the trace distance contraction coefficient of quantum channels. Provides bounds under channel composition that single scalar metrics cannot.
Topological control framework for quantum chaos diagnostics using OTOCs, spectral statistics, and information scrambling. Bridges network topology with information-theoretic, operator, and spectral diagnostics for quantum many-body systems.
Ravine analysis methodology for variational quantum algorithm (VQA) optimization using nudged elastic band (NEB) algorithm from theoretical chemistry to find low-cost paths connecting local minima in quantum cost landscapes, enabling ensemble prediction…
Quantum-inspired tensor network methodology for modeling order-dependent emotional memory in children, achieving 77.98% accuracy by incorporating valence into tensor factorization. Based on arXiv:2606.28470.
CLAIMSTAB-QC framework for auditing empirical comparisons in quantum software papers. Records baselines, metrics, and evidence; locks comparison design before outcomes; classifies reported directions as Sustained, Unresolved, or Reversed within locked audit…
Quantum entangled PET imaging using Compton events methodology - leveraging quantum entanglement for improved clinical PET imaging with better coincidence detection and spatial resolution at clinically relevant activities.
Quantum entanglement-based PET imaging methodology using Compton scattering polarization correlations for novel tissue biomarkers
Stable Self-Modulating Quantum Fast-Weight Programmers (QFWPs) with bounded memory gates for quantum sequence modeling. Prevents long-sequence divergence via sign-preserving tanh gates on recurrent memory branch. Based on arXiv:2607.02363.
Quantum Fast-Weight Programmers with bounded memory gates for stable quantum sequence modeling. Input-dependent gates for fast-weight updates with sign-preserving tanh stabilization for long-sequence regimes. Applicable to quantum dynamics forecasting and…
Controlled benchmark methodology for quantum vs classical generators in medical imaging with matched parameters (arXiv:2606.18970)
Gate-efficient quantum medical image encoding using Fourier-based methods. Reduces gates by 4x compared to pixel count. Covers two compression techniques for large-scale medical imaging with negligible quality loss.
Quantum-inspired computational framework for harmonic decision-making in music generation. Combines interference-based harmonization with classical tonal optimization to explore multiple chord sequences in parallel. Use when: music generation with…
Post-quantum cryptography and edge-native security patterns for Internet of Medical Things (IoMT) systems, including federated learning, PQC migration, and Kubernetes-based orchestration.
Hybrid quantum-classical surrogate model for Lattice Boltzmann Method (LBM) collision dynamics. Uses parameterized quantum circuits with data re-uploading to implement partial Fourier series, recovering complete BGK collision dynamics across full physically…
Logical spectroscopy methodology for constructing addressable conjugate bases in Abelian lifted-product quantum LDPC codes using CRT decomposition of group algebras.